Results arrive, revenue grows, and the share price falls. Lin opens her report to soften the sentence “Growth will persist.” If she overwrites the first version, she will later struggle to distinguish timely correction from rewriting what she supposedly believed before the result.
Research updates preserve the past while allowing the present judgment to change. Follow a membership-fee hypothesis through a new disclosure, then use an independent small model to distinguish a changed fact, a changed valuation, and a changed action.
What you will be able to do
- Compare an original hypothesis with new evidence, distinguishing expected variation, structural changes, and definition changes.
- Identify whether the revision affects a fact, assumption, valuation, or action, with a recorded reason.
- Produce a change note with conditions for pausing, abandoning, or continuing the investigation.
Save the original question first
C1's observation cutoffs remain December 31, 2024, and April 30, 2025. This chapter is a retrospective historical teaching reconstruction. The author knows later information, so it is not a genuinely unknown forward test. Readers who already know the outcome should also label their work retrospective reconstruction.
At the first observation point, one available document is Costco's 10-K for the fiscal year ended September 1, 2024. The SEC index lists its Filing Date as October 9, 2024. The report describes a U.S. and Canadian membership-fee increase effective September 1, 2024, with fee revenue deferred and recognized ratably over the one-year membership period. [1]
This is part of the evidence available at that point, not the complete initial packet. It supports a working hypothesis: the fee increase may gradually raise recognized membership-fee revenue, with the actual effect also depending on member numbers, renewals, exchange rates, and recognition timing. It does not establish that all revenue next quarter must rise proportionately with the fee.
Pause and save one page before continuing. Record that hypothesis, its source, unknowns, fields to check next, and evidence that would weaken it. Do not estimate Costco's actual target price. If you write the page only after reading the later results, mark it retrospective reconstruction rather than backdating it as an earlier judgment.
Check the date before the direction of the result
At the second observation point, Costco's 10-Q for the quarter ended February 16, 2025, is available. Its SEC index lists a Filing Date of March 13, 2025, between the two cutoffs. [2] That is the filing record for this document. It does not establish that every item first became public that day; actual event research also needs earlier announcements and information-availability times.
Compare the reported quarters of twelve weeks each, rather than a twelve-week amount with a twenty-four-week cumulative amount. Membership-fee revenue rose from $1,111 million to $1,193 million. Calculating from those disclosed amounts gives approximately 7.38% growth; the report describes it as 7% using a whole-number percentage.
The first update is a fact: recognized membership-fee revenue increased over the comparable period. Management's discussion includes new members and higher fees as drivers, with foreign exchange also affecting the result. [2] Neither the entire increase nor every related hypothesis can therefore be attributed to the fee increase alone.
Renewal rates need their own clock. The report describes a trailing calculation covering renewals seven to eighteen months before the reporting date. That information is useful without measuring every member's immediate reaction after the fee change. A familiar metric name does not establish that its time window answers the current question.
A defensible update is: “Higher membership-fee revenue is consistent with part of the original hypothesis. Several drivers contribute, and the observation window still limits assessment of short-term renewal effects.” That is more precise than “Growth proves I was right,” while avoiding the opposite error that incomplete confirmation automatically disproves the hypothesis.
Send each change back to the right step
Expected variation lies within conditions the original analysis already allowed. If a supported forecast range contains the result, inspect whether the drivers also behaved as expected. A number can land in the range for reasons very different from those predicted. Being inside the interval does not complete the explanation.
A structural change can break a key condition: a major customer permanently leaves, a product is replaced, contractual rights change, or financing becomes infeasible. That can require rebuilding the business, cash, and valuation case, rather than trimming the next quarter's growth estimate while keeping the rest unchanged.
A definition change returns first to the data. Reclassification, changed consolidation scope, different accounting periods, or new metric definitions can undermine comparability. Bridge old and new fields before judging whether the economics changed. A different measuring instrument is not automatically evidence of improved operations.
Resolve source conflicts carefully too. If a news item disagrees with a statement, check the original units and periods. If a later filing amends an earlier disclosure, retain both versions and when each was available. EDGAR uses the “/A” suffix to identify amendments, but you must read the filing to establish what changed. [3]
Do not overwrite historical files with today's corrected values and assume earlier decisions used them. Current analysis seeks the best available facts now; historical evaluation also needs the version knowable then. Separate columns let both questions be investigated without quietly substituting one for the other.
Four layers prevent a one-sentence leap
First record the fact: what the new material says, its source, date, and scope. Then update the hypothesis: which explanation it supports or challenges, and which alternatives remain. Third, identify affected revenue, cost, investment, or risk assumptions in the valuation. Only then record the action, such as investigating, pausing, abandoning a premise, or reevaluating a position under relevant constraints.
Abandoning a hypothesis does not always mean immediately selling. You may own no position, or first need to check security rights, price, and execution. Conversely, a supported operating hypothesis can coexist with a need to alter exposure because household spending, concentration, or valuation changes. Record those reasons separately instead of presenting them all as judgments about company quality.
Make the downstream change visible in a small model
Leave Costco entirely and use Jiangsen Equipment, a fictional independent example. Model money is in units of $10,000. One million common shares remain unchanged, with no other claims. The model values the next complete year's forecast common-shareholder profit at today's valuation date; none of the prices are actual company quotations.
The first version assumes profit of 100 and a P/E reference of 15. Simplified equity value is 1,500 model units, or $15 a share. New material leads the researcher to revise comparable sustainable forecast profit to 80. Keeping P/E at 15 gives equity value of 1,200 units and $12 a share. The controlled example isolates profit; real risk and multiples need review too.
This changes a forecast, not an already realized historical profit figure. Record the evidence and reasoning for moving from 100 to 80, why it is more than temporary recognition timing, and why the multiple is provisionally unchanged. Without sufficient support, label the result a stress scenario rather than a newly established valuation.
Now suppose the investor previously paid $10 and the current supplied price is $9, ignoring costs. The 10% loss relative to cost differs from the valuation revision from $15 to $12. Price falling below the purchase price does not prove the operating premise failed. Nor does $9 being below $12 automatically justify adding to the position.
That decision still depends on the reliability of 80, the multiple, financial loss capacity, and implementation conditions. If new information invalidates a going-concern earnings valuation, pause that method. Retaining $12 solely as reassurance would hide the model limitation rather than resolve it.
A failed business premise and a price stop do different jobs
Here, a business-based exit condition means a key operating assumption no longer holds and the old conclusion should no longer be used. It is a research rule, not a brokerage order or fixed percentage decline. A price stop instead responds mainly to market price as part of a risk rule or order arrangement.
A stop-market trigger does not guarantee its execution price; gaps and liquidity can alter the outcome. [4] Neither a business condition nor a price trigger should become a promise about maximum loss. The purpose and evidence for each condition belong beside its description.
You might return to the business hypothesis after reliable customer-contract termination evidence, reassess funding and avoid new positions when cash timing no longer matches obligations, or review execution when an existing price rule triggers. The specific thresholds need goals and constraints; this chapter provides no universal exit level.
Persistence needs evidence as much as correction does. If unfavorable news does not reach a key assumption, explain why maintaining the view is justified. “Long-term investing” alone supplies no such explanation. Equally, a falling share price is not a complete argument that every prior judgment was wrong.
Make an update before checking the answers
Why does higher quarterly membership-fee revenue fail to establish, by itself, that higher fees had no renewal effect? In Jiangsen's model, someone changes forecast profit from 100 to 80 but keeps the report's $15 conclusion. Which layer needs correction? Write your answers first.
The first needs both multiple revenue drivers and the renewal metric's delayed observation window. For Jiangsen, valuation must update to $12 while retaining the old $15 and revision reasons. If the profit change is only a temporary stress case, the conclusion must carry that scenario label rather than mix first-version assumptions with the new result.
Change the evidence rather than the economics
In another independent teaching scenario, Jiangsen combines two revenue classifications. Total revenue is unchanged, with no new contract or cost developments. Meanwhile, the share price rises from $10 to $11. Someone increases forecast profit from 100 to 110, saying price has confirmed earnings. What should happen first?
Build the classification bridge and record the reporting change. The supplied material adds no support for higher operating profit, and a stock price does not replace a forecast input. Retain the working forecast if still supported, record unknowns and the price movement, and revisit valuation and risks. Do not insert 110 merely because the quote is $11.
For C1-B Target, begin with your own first-version question and documents actually available at the two dates, then identify which layer changed. Without an original record, label the work retrospective reconstruction. Do not import Costco's fee mechanism. Target's fiscal-2024 10-K was filed March 12, 2025: it can enter T2, not T1, and a C1-R usage label cannot change that timing.
Submit a T04 change note
Put the first report beside new material. For each core claim, record the original, new evidence and time, comparability check, affected layer, retained or revised conclusion, missing evidence, next action, and review condition. “Keep watching” is incomplete unless you specify what to watch.
Preserve the first report and every change, linking the new version to the old one and its sources. For corrected historical data, retain a separate as-known-then value. Mark each core claim retained, revised, or suspended with reasons. New evidence can legitimately leave every original conclusion unchanged; do not invent a reversal merely to complete the exercise. Use the separate 100-to-80 profit example to demonstrate a genuine valuation update.
Cover the later stock price and ask whether the revision reasons still stand. Then cover the purchase price and check whether goals, evidence, and risks still explain the proposed action. The record is the company-research deliverable; the synthetic exercises do not establish completion of real Costco research.
Change today's judgment without erasing yesterday's
Check new evidence for comparability and availability, then distinguish facts, assumptions, valuation, and action. Structural business changes, reporting changes, and market-price changes need different support. A favorable report or falling quote cannot stand in for every step.
Preserve reasons for both revision and retention, with reviewable conditions for continuing, pausing, or exiting. Explaining why you changed, and why you sometimes did not, turns follow-up into research rather than a rewritten story about the outcome.
References
- Costco fiscal-2024 10-K, Item 7 Membership Fees and Note 1; SEC filing index. This document predates the first cutoff without representing a complete initial packet.
- Costco 10-Q for the quarter ended February 16, 2025, Item 2 Membership Fees; SEC filing index. Preserve Filing Date separately from Accepted; a reporting period-end is not a publication date.
- Investor.gov: Using EDGAR to Research Investments. Document types and amendment identifiers.
- Investor.gov: Types of Orders. Stop triggers and fill prices. Sources checked October 30, 2026.
This is historical teaching reconstruction plus independent synthetic exercises. AI assisted with writing and checks; independent AI full-text and bilingual review is complete; human financial review and target-reader testing remain pending. No real-company trade recommendation is supplied.